Each year numerous automobile accidents are caused by driver distractions, and many of the distractions are visual in nature. For this reason, there has been interest in developing a driver monitoring system for determining if the driver is paying attention to the forward field-of-view. This information can be used to issue an alert if the driver's attention is directed away from the road too long or too often, and possibly to belay other warnings (such as collision-avoidance warnings) if the driver is paying attention to the forward field-of-view. An example of such a monitoring system is Delphi Corporation's Driver State Monitor, which processes a video image of the driver's face to detect and track the driver's eyes for assessing eye gaze. However, detection of facial features such as eyes can be hampered by various kinds of obstructions (including sunglasses) disposed between the video imager and the driver's face. Moreover the distance between the video imager and the driver's face can vary considerably from driver to driver, and it can be difficult to provide adequate controlled illumination of the driver's face. While these drawbacks can be satisfactorily addressed to a large extent by sophisticated processing of the video data, the system cost is frequently too high for most production vehicles due to the combined cost of the imager, optics and signal processor. Accordingly, what is needed is a more cost-effective and yet reliable way of assessing driver head pose.